J. Am. Chem. SOC. 1994,116, 8105-8111
8105
Acetylene a-Coordination, Slippage to a-Coordination, and 1,ZHydrogen Migration Taking Place on a Transition Metal. The Case of a Ru(I1) Complex As Studied by Experiment and ab Initio Molecular Orbital Simulations Yasuo Wakatsuki,'*+Nobuaki Koga,* Hiroshi Yamazaki,g and Keiji Morokumal Contribution from The Institute of Physical and Chemical Research (RIKEN), Wako, Saitama 351-01. Japan, Department of Chemistry, School of Informatics and Sciences, Nagoya University, Chikusa, Nagoya 464-01, Japan, Department of Chemistry, Faculty of Science and Engineering, Chuo University, Kasuga, Bunkyo, Tokyo I 12, Japan, and Cherry L. Emerson Center for Scientific Computing and Department of Chemistry, Emory University, Atlanta, Georgia 30322 Received April 15, 1994"
Abstract: The behavior of an acetylene molecule in the coordination sphere of transition metals has been probed by the reactions of RuX2(PPh3)3 (X = C1, Br) with tert-butylacetylene to give vinylidene complexes of the formula RuXz(PPhl)2(C=CHtBu). IR and NMR data have indicated that the initial product of this reaction is a mixture of two complexes each of which has a vinylidene unit and nonequivalent cis-bis(phosphine) ligands. In solution, these kinetic products gradually isomerize to the final trans-bis(phosphine) complex. The structure of this five-coordinated and thermodynamically stable complex (X = Br) was determined by X-ray crystallographic analysis to have a quasi trigonal-bipyramidal conformation with the two phosphines occupying axial positions. The potential surface for the transformation of coordinated acetylene to vinylidene was calculated by the ab initio molecular orbital method. The primary process was concluded to be a slippage of the $-CC coordinated alkyne to the q2-CH coordinated complex via a transition state with an VI-acetylene and a side-on acetylene. The 92-CH complex undergoes 1,Zhydrogen migration within the acetylene unit, whose transition state is the highest point of the whole process, giving finally the thermodynamically metastable vinylidene complex with a cis-bis(phosphine). The isomeric vinylidene ruthenium(I1) complex with trans-bis(phosphine) has been calculated to be the final product and thermodynamically most stable form of this reaction system. The role of the metal in the present rearrangement is discussed on the basis of the localized molecular orbital analysis of the key intermediates.
The electronic character and reactivity of organic compounds change, often dramatically, when they interact with transition metals. Since this phenomenon is strongly related to many catalyzed processes, clarification or understanding of this basic problem is an important topic for organometallic as well as theoretical chemists. The relative stability of acetylene and its isomeric vinylidene form provides a good and simple example: free vinylidene, :C=CH2, has been the subject of a number of theoretical and physicochemical studies1 and is proved nowadays to be ca.44-47 kcal/mol less stable than acetylene, H W H . 1 - 3 In contrast, 1-alkyne to vinylidene tautomerization in the coordinationsphere of a transition metal has proved to be a useful entry into vinylidenecomplexes,apparently with vinylidene being the more stable form in many transition-metal complexes.4.5Such Institute of Physical and Chemical Research (RIKEN). t Nagoya University. i Chuo University. I Emory University. * Abstract published in Aduance ACS Absrracrs, August 1, 1994. (1)(a) Krishnan, R.; Frisch, M. J.; Pople, J. A.; Schleyer, P. v. R. Chem. Phys. Lerr. 1981,79,408 and references cited therein. (b) Osamura, Y.; Schaefer, H. F.; Gray, S.K.; Miller, W. H. J . Am. Chem. Soc. 1981,103, 1904. (c) Carrington, T.; Hubbard, L. M; Schaefer, H. F.; Miller, W. H. J . Chem. Phys. 1984,80,4347.(d) Gallo, M. M.; Hamilton, T. P.; Schaefer, H. F. J. Am. Chem. Soc. 1990,112,8714. (e) Petersson, G.A.; Tensfeldt, T.G.; Montgomery, J.A., Jr.J.Am. Chem.Sm. 1992,114,6133.(f) Paldus, J.; Cizek, J.; Jeziorski, B. J. Chem. Phys. 1989,90,4356.(g) Jensen, J. H.; Morokuma, K.; Gordon, M. S . J . Chem. Phys. 1994,100, 1981. (2) (a) Ervin, K. M.; Ho, J.; Lineberger, W. C.J. Chem. Phys. 1989,91, 5974. (b) Ervin, K. M.; Gronert, S.;Barlow, S.E.;Gilles, M. K.; Harrison, A. G.; Bierbaum, V. M.; DePuy, C. H.; Lineberger, W. C.; Ellison, G. B. J . Am. Chem. Soc. 1990,112,5750. (3) (a) Chen, Y.;Jonas, D. M.; Hamilton, C. E.; Green, P. G.; Kinsey, J. L.; Field, R. W. Ber. Bunsen-Ges. Phys. Chem. 1988,92,329.(b) Chen, Y.; Jonas, D. M.; Kinsey, J. L.; Field, R. W. J . Chem. Phys. 1989,91,3976.
isomerication on transition metal sometimes plays a key role in catalytic C - C coupling reactions: coupling of metal-alkynyl and alkyne carbons to form a metal-bound butenynyl group is considered to proceed via prior isomerization of the 1-alkyneinto the vinylidene form (eq 1). Recently, Bianchini and co-workers and we have shown that such C-C coupling indeed takes place in Ru(I1)-catalyzed dimerization of 1-alkynes to Z-1,Cdisubstituted enynes6and Z-1,Cdisubstituted butatrienes? respectively. U
(4)(a) Bruce, M.I. Chem. Rev.1991,91,197and references therein. (b) Werner, H. Angew. Chem., Int. Ed. Engl. 1990,29, 1077 and references therein. (5) H6hn,A.; Werner, H.J. Organomet. Chem. 1990,382,255.(b) Werner, I+.; Rappert, T.; Wolf, J. lsr. J. Chem. 1990,30,377.(c) Werner, H.; Stahl, S.;Kohlmann, W. J. organomet. Chem. 1991, 409, 285. (d) Knaup, W.; Werner, H. J. Organomet. Chem. 1991,411,471.(e) Werner, H.;H6hn, A,; Schulz, M. J. Chem. Soc, Dalton Trans. 1991, 777. (f) SchHfer, M.; Wolf, J.; Werner, H. J . Chem. Soc., Chem. Commun. 1991,1341. (g) Schneider, D.; Werner, H. Angew. Chem. 1991,103,710.(h) Werner, H.; Dirnberger, T.; HBhn, A. Chem. Ber. 1991,124,1957. (i) Werner, H.; Weinhand, R.; Knaup, W. Organometallics1991,10,3967.(j)Wemer,H.;Stark,A.;Schulz, M.; Wolf, J. Organometallics 1992,II,1126. (k) Rappert, T.; 0.; Mahr. N.; Wolf, J.; Werner, H. Organomeralllcs 1992,11,4156.(1) Werner, H.; Weber, B.; NBrnberg, 0.;Wolf, J. Angew. Chem., Inr. Ed. Engl. 1992,31,1025.(m) Nakanishi, S.;Gaia, K.; Uchiyama, S.;Otsuji, Y.Bull. Chem. Soc. Jpn. 1992,65,2560.(n) Haquette, P.;Pirio, N.;Touchard,D.;Toupet,L.; Dixneuf, P. H. J. Chem. SOC.,Chem. Commun. 1993,163. (6) Bianchini, C.; Peruuini, M.; Zanobini, F.; Frcdiani, P.; Albiati, A. J . Am. Chem. Soc. 1991, 113,5453.
0002-7863/94/ 1516-8105$04.50/0 0 1994 American Chemical Society
8106 J. Am. Chem. SOC.,Vol. 116. No. 18, 1994 Regarding the mechanism of the conversion of coordinated 1-alkynes to metal-bound vinylidenes, Antonova and co-workers have suggested that oxidative addition of the alkyne to give an alkynyl(hydrid0)metal intermediate and subsequent 1,3-shift of the hydride from the metal to the @-carbonof the alkynyl might give the vinylidene complex.* However, the extended Hiickel calculations on Mn(CSHs)(C0)2(C2H2)have suggested that the concerted intramolecular 1,3-hydrogen shift of the alkynyl(hydrido)metalcomplex would have a very high activationenergy and that a direct 1,2-hydrogen shift from the a- to @-carbon might be plausible? A series of works by Werner and co-workers appears to have provided experimental evidence to judge which of these pathways is more likely: they have shown that squareplanar s2-alkynecomplexes trans-[MC1(R-CH)(PiPr&] (M = Rh, Ir) rearrange to square-pyramidal alkynyl(hydrid0) complexes MCl(PiPr3)z(H)(C=CR), which on warming further isomerize to the vinylidene complexes trans-[MCl(C=CHR)(PiPr3)2].lo A little different from Antonova's proposal, the possibility of an intermolecular hydrogen shift between two alkynyl(hydrid0)metalmolecules has been pointed out.'OC More recently, Bianchini and co-workers havecarried out kinetic studies on the rearrangement of a cationic alkynyl(hydrido)cobalt to its vinylidene form and proposed that it proceeds via a dissociative (nonconcerted) intramolecular 1,3-proton shift." Obviously, too many possibilities exist once acetylene interacts with transition-metal complexes. Along with a diversity of configurationsaround thecentral metal, this probem is a challenge to theorists. Ab initio MO calculations on the acetylenevinylidene rearrangement under the influenceof a non-transitionmetal atom have been reported,12 but to our knowledge no ab initio theoretical report has appeared with transition-metal complexes. Seeking to model the key step (eq 1) in our catalytic dimerization described above,' we have found that a ubiquitous ruthenium complex, RuXz(PPh3)s (X = C1 or Br), reacts readily with tert-butylacetylene to give a stable vinylidene complex (eq 2). Since reaction 2 seemed rather simple and suitable for
Wakatsuki et al.
u /I Figure 1. Molecular structure for 2f with atomic-numbering scheme.
phosphines at axial positions but was deformed considerably toward square pyramidal with the vinylidene C, at an apical position, the Br-Ru-Br angle being 145.6(1)O. Though fivecoordinated d6 metal complexes have been calculated to adopt a square-pyramidal structure rather than a trigonal-bipyramidal one, the energy difference is small.l3 The large deviation of the Br-Ru-Br angle from linearity, with deformation toward the trigonal bipyramid, should be favored when steric repulsion between the Br's and the vinylidene moiety exists. The M-C-C RuX,(PPh,), HCEC'BU angles found in vinylidene complexes often deviate from linearity, RuX,(PPh,),(C=CH'Bu) PPh, (2) but the present value of 161(2)O is one of the largest.4 This is most likely due to severe steric repulsion between the phenyl mechanistic investigations, we decided to explore this system in rings of PPh3 and the bulky tBu group at Cg, as easily seen from detail with both experiment and ab initio theoretical calculations. Figure 1b. NMR and X-ray analysis were used to determine the orientation When the reaction of RuC12(PPh3)3 with H"Bu was of ligands around the central metal while theoretical calculations stopped at an early stage, e.g. 15 min, by evaporating the excess were successfully applied to elucidate the transition states and alkyne and the solvent, a fine, red-brown powder was obtained. unstable key forms of the C2 unit under such circumstances. It showed an IR absorption at 1638 cm-l, the typical v(C==C) region for a vinylidene ligand, but no peaks at all at 1700-2000 Results and Discussion cm-l, the region for +coordinated alkyne or a-alkynyl. The 1H (1) ReactionsofRuCI~(PPh3)~andRuBr~(PPh3)3withH~- NMR spectrum in C6D6 showed the 'Bu resonance at 6 0.91 (broads) and a proton absorption at 2.41 (broad t, J = ca. 4 Hz). *Bu.The complexes RuXz(PPh3)3 (X = C1, Br) react with excess When theNMRsample wasallowed to stand at room temperature, HC=CtBu in benzene over 24 h at room temperature to give the intensities of these peaks gradually decreased while the peaks vinylidene complexes RuX2(PPh3)2(C=CHtBu) (lf, X = C1; Zf, dueto lfat0.79 (s,tBu) and 3.87 ( t , J = 4,4Hz,-CH)increased. X = Br), whose spectral data have been previously reported.' The change was completed within 30 h. The 31P NMR of the The X-ray structural analysis of 2f (Figure la) showed that initial brown-red powder in C& revealed that this was a mixture the molecule had a trigonal-bipyramidal structure with two of almost equal amounts of two isomers, each of which had two (7) Wakatsuki, Y.; Yamazaki, H.; Kumegawa, N.; Satoh, T.; Satoh, J. Y. nonequivalent phosphines: the spectrum exhibited two AB-type J . Am. Chem. Soc. 1991. 113.9604. quartets centered at 6 46.5 ( J = 38 Hz) and 34.4 (J = 25 Hz) (8) (a)Nesmeyanov, A.N.;Aleksandrov,G.G.;Antonova,A. B.;Anisimov, K. N.; Kolobova, N. E.; Struchkov, Yu. T. J. Organomet. Chem. 1976,110, with almost equal intensities. A small amount (ca. 5%) of a C36. (b) Antonova, A. B.; Kolobova, N. E.; Petrovsky, P. V.; Lokshin, B. V.; singlet at 6 27.2 due to If was also observed, besides a singlet for Obezyuk, N. S. J. Organomet. Chem. 1977,137, 55. free PPh3 liberated as indicated in eq 2. When the NMR sample (9) Silvestre, J.; Hoffmann, R. Helu. Chim. Acta 1985, 68, 1461. (10) (a) Wolf, J.; Werner, H.; Serhadli, 0.;Ziegler, M. L. Angew. Chem., was allowed to stand at room temperature, the two quartets Int. Ed. Engl. 1983, 22,414. (b) Hbhn, A.; Otto, H.; Dziallae, M.; Werner, decreased as the singlet of If at 27.2 increased with a rate almost H. J . Chem. SOC.,Chem. Commun. 1987, 852. (c) Werner, H.; Alonso, F. equal to that observed for the change of the 1H NMR spectra. J. G.; Otto, H.; Wolf, J. 2.Naturforsch. 1988, 436, 722. (d) Werner, H.; Brekau, U. 2.Naturforsch. 1989, 446, 1438. (e) Hbhn, A.; Werner, H. J . On the basis of these spectral data, we have concluded that the
+
-
+
Organomet. Chem. 1990, 382, 255. (1 1) Bianchini, C.; Peruzzini, M.; Vacca, A.; Zanobini, F. Organometallics 1991, 10, 3697. (12) Sakai, S.;Morokuma, K. J. Phys. Chem. 1987, 91, 3661.
(13) (a) Elian,M.;Hoffmann,R.Inorg. Chem. 1975,14,1058. (b) Daniel, C.; Koga, N.; Han, J.; Fu, X. Y.; Morokuma, K. J . Am. Chem. SOC.1988, 110,3773.
Behavior of Acetylene in a Coordination Sphere
J. Am. Chem. SOC.,Vol. 116, No. 18, 1994 8107 1.257 1.070
Scheme 1
082
1.17
PPh3
Pbh,
lb
PPh3
It PPh,
if
1.215 1.061
le
initially formed brown-red powder consists of two isomers, Id and le, depicted in Scheme 1. The lH NMR absorptions of the tert-butyl and vinylidene protons must have coincided between Figure 2. Comparison of the potential energy for the isomerization of both isomers, since they are remote enough from the central metal, acetylene to vinylidene in the free state (upper line) and in the rutheniumand the peaks are broad besides. Efforts to detect v2-alkyne (11) complex (lower line). The units are A for atomic distances and complex lb have been unsuccessful: shorterreaction times resulted kcal/mol for energies. in a lower yield of the vinylidene complex with more unreacted material. C2H2 moiety. To refine the energetics of the thus optimized The reaction route we propose is illustrated in Scheme 1 and structures, the MP2 energies were recalculated using the 16is fully consistentwith the known structure of the starting RuC12valence-electron effective core potential16bfor Ru, which takes (PPh3)3. The X-ray structural analysis of R u C ~ ~ ( P has P ~ ~ ) ~explicitly into account the outermost core orbitals, with the shown that it has a slightly distorted square-pyramidal configu[3~3p2d]/(5s5p4d)basis functions. ration.14 In the solid state, the metal weakly interacts with one Before calculating the complexes, we examined the rearrangeof the phenyl hydrogens (Ru-H = 2.59 A), but in solution it ment of the free C2H2 unit. Many theortical studies have already should open the coordination site. Though we have been unable focused on the isomerization of free vinylidene (:C=CH2) to to detect the $-alkyne complex, it is reasonable to assume that acetylene (HC=CH),I a reverse process of what we are coordination of the alkyne through the triple bond is involved in considering here. Petersson, Tensfeldt, and Montgomery have the present reaction sequence, since the alkyne ?r-complex is a recently provided a comparison of the various levels of ab initio thermodynamically stable species but less stable than the final theory for this isomerization reaction, including HF, MP2, and vinylidene complex (vide infra). The bulky alkyne attacking this CCSD (coupled cluster single and double excitations) methods site will displace one of the neighboring triphenylphosphinesto with STO-3G, 6-31G**, and 6-311 G** basis sets.le The form an q2-CCalkynecomplex I b with cis-bis(phosphine)swhich calculated geometriesfor vinylidene, acetylene,and the transition quickly isomerizes to the vinylidene complex Id with the same state have been found to be fairly insensitive to the level of theory metal configuration and with the tert-butyl group occupying the in spit of the flat potential surface between the vinylidene and least crowded position. Our ab initio MO calculationson RuC12the transition state. In contrast, the height of the calculated (PH&(C=CH2) have suggested that a rotational isomer l e is barrier was very sensitiveto the level of theory: Too large barriers electronically favorable (vide infra). But the bulky tert-butyl were obtained unless electron correlation was induced; e.g., HF group present in the actual reactant should prefer Id over l e in calculation with a large basis set gave a barrier for vinylidene to view of steric effects, thus equalizingthe total energiesof the two acetylene isomerization of 12.1 kcal/mol with an isomerization isomers giving a 1 to 1 mixture. Complex l e will slowly isomerize energy of -35.2 kcal/mol. Only a sophisticated method, the to the final product If, which has been calculated to be the most quadratic configurationinteraction (QCI) calculationswith large stable form as described in the next section. basis sets could reproduce the experimentallyobserved values of (2) Molecular Orbital Calculations on the Intramolecular the reaction barrier (2 kcal/mo12) and the energy of isomerization Rearrangement of RuC12(PH3)2(HC=CH) to RuCIz(PH3)2(-47.4 f 4.0 kcal/mol,2 -44.4 f 0.3 kcal/mo13). The results by (C-2). By means of ab initio molecular orbital calculations MP2/6-3 lG* calculations have been found to be qualitatively (GAUSSIAN 92 programI5), we examined if there exists a satisfactory (barrier 1.07,isomerization energy-48.7 kcal/mol).'e reasonable path for the concerted intramolecular rearrangement It appears, therefore, reasonable in view of economy to optimize of the ?r-coordinated acetylene complex to the vinylidene the geometry by the MP2 method. The geometries and energies ruthenium complex. Throughout the calculations, except for of free acetylene, vinylidene, and the transition state calculated complex G with a d4 metal configuration (vide infra), the by this methodology and with the basis sets we have employed geometries of the stationary points were optimized at the MP2 as described above are shown by the upper line in Figure 2. The (electron correlation by second-order Maller-Plesset perturbaenergy difference (0.1 kcal/mol) between the vinylidene form tion) level using [2s2p2d]/(3s3p4d) and the eight-valence-electron and the transition state, and the total isomerizationenergy (-5 1.7 effective core potential of Hay and Wadtl6a for the metal, the kcal/mol) deviateby a few kcal/mol from the experimetnalvalues 4-31G basis set1' for Cl and P, the 3-21G set1*for the H of PH3, but are qualitatively, if not quantitatively, satisfactory. the 4-31G* basis set19for C, and the TZP set20for the H of the Performing calculations on the C2H2 unit with the R u ( P H ~ ) ~ (14) La Placa, S. J.; Ibers, J. A. Inorg. Chem. 1965, 4, 778. C12 moiety, we have found altogether six complexes, A-F, and
+
(15) Frisch, M. J.; Trucks, G. W.; Head-Gordon, M.; Gill, P. M. W.; Wong, M. W.; Foresman, J. B.; Johnson, B. G.; Schlegel, H. B.; Robb, M. A,; Replogle, E. S.;Gomperts, R.; Andres, J. L.; Raghavachari, K.; Binkley, J. S.;Gonzalez,C.; Martin, R. L.; Fox, D. J.; Defrees, D. J.; Baker, J.; Stewart, J. J. P.; Pople, J. A. GAUSSIAN92; Gaussian, Inc.: Pittsburgh, PA, 1992. (16) (a) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985,82,270. (b) Hay, P. J.; Wadt, W. R. J . Chem. Phys. 1985,82, 299. (17) Hehre, W.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972,56,2257.
(18) Binkey, J. S.;Pople, J. A.; Hehre, W. J. J . Am. Chem. SOC.1980,202, 939. (19) (a) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971, 54,724. (b) Hariharan, P. C.; Pople, J. A. Theor. Chim.Acta 1973,28,213. (20) (a) Dunning, T. H. J. Chem. Phys. 1971, 55, 716. (b) Dunning, T. H.; Hay, P. J. Modern Theoretical Chemistry; Plenum: New York, 1977; VOl. 2, p 1.
Wakatsuki et al.
8108 J. Am. Chem. Soc., Vol. 116, No. 18, 1994 H
A
C P-Ru-P 94.9 CI-Ru-CI 161.9 152.7 C-C-H
2.259
P-Ru-P CI-Ru-CI Ru-C-C C-C-H,
90.6 175.3 154.9 147.5
Cp-C,-H,
91.4 174.1 169.5 173.1 00.5
H
B
P-Ru-P 91.5 CI-Ru-CI 173.9 C-C-H,, 160.5 C-C-HhWn152.9
D
/$
Y
P-Ru-P 90.0 CI-Ru-CI 172.4 P-Ru-C 00.3 Ru-C-C 179.9
P-Ru-P 90.3 CI-Ru-CI 174.4 C-C-Ha 170.2 P-Ru-C 164.0
Y
E
2.240 P-Ru-P CI-Ru-CI P-Ru-C Ru-C-C
90.0 165.8 89.9 175.4
12.431
F
CI-Ru-CI 150.5
G
2.584
P-Ru-P 93.3 CI-Ru-CI 179.0 P-Ru-C 94.6 04.4 H-Ru-C
Figure 3. Calculated structures (in A and deg) of RuClz(PH&(H-CH),
R u C ~ ~ ( P H ~ ) ~ ( Cintermediates, ~H~), and the transition states.
two transition states, [TSl]and [TS2],whoseoptimized structures are shown in Figure 3 while their relative energiesare summarized in Figure 4. In the geometry optimizations, Cb symmetry was assumed for the final product F,and C, symmetry, for the others.
Complexes A and B are rotational isomers of the $-CC acetylene complex with cis-bis(phosphine)and trans-dichloride. Isomer B is less stable than A by 13.5 kcal/mol and is likely to be the transition state for this rotation. We could not find an
J . Am. Chem. SOC.,Vol. 116, No.18, 1994 8109
Behavior of Acetylene in a Coordination Sphere
F -59.6
(-70.3)
Figure 4. Energy diagram of RuC12(PH&(HC--CH), RuC12(PH3)2(C=CH2), intermediates, and the transition states. Values in parentheses are energies calculated using’the eight-valence-electron effective core potential for Ru.1b
acetylenecomplex with trans-bis(phosphine)and trans-dichloride probably because of steric reasons. On the vinylidene complex side, however, such a conformation gives the most stable form F. As described in section 1, the primary vinylidene product must have two nonequivalent phosphines. Of the three differentsquarepyramidal geometries of vinylidene complexes which have nonequivalent cis-phosphines, forms D and E (cis-phosphinesand trans-chlorides) are found to be energetically favorable (the other two have cis-phosphines and cis-chlorides). The rotational isomer D, rotation around the metal-carbon bond by 90°,was calculated to be 9.6 kcal/mol higher in energy than E. Similarly, the rotational isomer of F has been calculated to be more unstable than F by 11.4 kcal/mol, but it is not shown in Figure 4 for simplicity. These less stable rotational isomers are probably the transition states for the vinylidene rotation around the metalcarbon bond. The more stablecharacterof the $-alkyne complexA compared to its isomer B,as well as the stabilityof thevinylidene Ecompared to the rotational isomer D, may be attributed to the efficiency of back-donation from the metal d r orbitals. The importance of back-bonding in low-valent transition-metal alkene (dr-7r*)21 and vinylidene (dr-empty p)22complexes has been established. Calculations on the RuC12(PH& fragment, which has transdichloride and cis-bis(phosphine) ligands, have shown that the d, orbital lies at a much higher energy level than does dxz.The energy differenceamounts to 0.09au, and this can be traced back to an antibonding interaction of d, with the C1 p orbitals as illustrated below. The d r orbital with higher orbital energy (d,) can more efficiently back-donate its electrons to the empty r* orbital of the incoming alkyne or to the empty p orbital of the vinylidene ligand.23 metal dn (occupied)
P“3
acetylene X* (empty)
vinylidene p (elvty)
J
z
t;
P The vinylidene complexes with cis-phosphines are much less stable than the trans-phosphine isomers, and we had to fix the
P-Ru-P angle at 90° in the geometry optimizations of D and E. Without this restriction, they collapsed into the conformations with trans-phosphines and thus they are not the stationary points. In the real reaction (Scheme l), however, corresponding cisisomers Id and l e were observed, probably owing to steric reasons. Steric repulsion between the triphenylphosphineligands and the tert-butyl group, as seen in Figure lb, should make the energy of If less stable than might be expected from the extremely stable calculation model F. It is very likely, therefore, that the cisconformers like D and E are in stationary states in reaction 2, which involves bulky substituents. With monosubstituted acetylenes, in particular taking into account the presence of the bulky tert-butyl group in our model reaction (Scheme l), the acetylene complex A should be far less stable than the energy level expected from Figure 4 due to steric repulsion between the substituent and chloride. We may assume reasonably that the acetylene complexB is much more important, since the substituent can have the most free empty space when it is attached to the upper carbon. The crucial hydrogen migration should occur on the way from B to D, and the migrating hydrogen must be that at the lower carbon in B. The transition state for the hydrogen migration step ([I&]) has been found, and the potential energy change in this process is compared in Figure 2 with that calculated for free acetylene. The C2H2 moiety in the [TSz] complex adopts a similar structure to that calculated for the transition state of free C2H2 but is situated a little more to the alkyne side; the Ca-HmimtingdiStanceS of the free and complexed transition states are 1.175 and 1.127 %i and those of CrHmigrating are 1.417 and 1.654A, respectively. The transition state in the complex is located early, reflecting the much smaller energy of isomerization. It is apparent from Figure 2 that, in the coordination sphere of the transition metal, the vinylidenelone pair can be stabilizedby coordination to the metal. The large endothermic transformation from free acetylene to vinylidene is now thermoneutral in the complexed form. The lowering of the transition-state energy in the complex by ca. 17 kcal/mol can also be attributed to coordination of the lone-pair electrons.24 Careful inspection of the path from B to [I&] has revealed further the presence of a q2-CH complex C and a small barrier (21) Kitaura, K.; Sakaki, S.;Morokuma, K. Znorg. Chem. 1981,20,2292. (22) Kostic, N. M.; Fenske, R. F. Organometallics 1982, I , 974. (23) The ?r*(LUMO)of theacetylene fragment in Aand empty p(LUM0) of the vinylidene fragment in D have energy levels higher than dw(HOMO) of the RuCIz(PH3)z fragment by 0.48 and 0.44 au, respectively.
Wakatsuki et al.
8110 J. Am. Chem. SOC., Vol. 116, No. 18, 1994
Table 1. Crystallographic Data for 2f C42HaBrZPZRu V = 1876A3 f w = 867.6 z=2 = 0.7107 A cryst class: monoclinic space group: P21 pc = 1.534 pcm-3 a = 13.139(3) A ~ ( M Ka) o = 6.65 cm-I b = 15.276(3) A cryst size: 0.33 X 0.23 X 0.12 mm3 c = 9.997(5) A R = 0.058 @ =110.74(3)’ R, = 0.067
A
-
Figure 5. Localized molecular orbitals in the plane of P-Ru-CH-CH lone pair; for intermediate C and the two transition states: left, C-H, right, (C-C)?r. [=I] between B and C (Figure 4). The rearrangement of B to C is a slippage of the +CC acetylene complex to the s2-CH
complex. In the transition state [TSl]for this slippage, the metal interacts dominantly with only C, and this interaction is very weak: the geometry of the C2H2 moiety deforms only slightly from the free acetylene and the Ru-C, distance is as long as 2.4 A (Figure 3). Parts a-c of Figure 5 compare shapes of CH u- and CC 7-orbitals in the complexes [TSI], C, and [TSz]. These MO contour diagrams clearly demonstrate that the C,-H, u-bond turns into the lone-pair orbital along the reaction coordinate while the in-plane CC *-orbital changes into the newly formed C r H , u-bond on going from [=I] to [TSz]. The p orbital a t C,, which has been a part of the in-plane 7-orbital, is left empty as the C r H a-bond is being formed, and it should end up as the “empty p” orbital of the vinylidene to accept back-donation from the metal. It may be inferred that the migrating hydrogen behaves as a naked proton rather than a hydride. The migrating hydrogen leaves behind the lone pair of electrons which was originally used for the Ca-H u-bond while it acceptsthe electrons, which originate from those of the in-plane CC 7-orbital, to form the new C r H u-bond. This view is in contrast to the concept proposed by Petersson et al.1e and Sakai et a1.12 for the free vinylidene acetylene isomerization where the lone pair of the vinylidene is correlated with a *-orbital of acetylene and the migrating hydrogen has hydride character.25 We conclude that the major role of Ru(I1) in this rearrangement reaction is to interact with the C-H a-bond first and then to promote its conversion to the
-
(24) As indicated in Figure 4, the energy level of [TSz] is still ca. 8 kcal/ mol higher than the dissociated state. This thermodynamics docs not mean that the complex C dissociates before it goes over [TSz]. We estimate that the transition energy needed for the dissociation of the acetylene unit from Ru is almost 50 kcal/mol, though we have not determined the exact transitionstate geometry. (25) Since acetylene kecps its interaction with Ru during the present reaction, the ambiguity of orbital transformation discussed by Petersson et al. is not met here. Our lone-pair LMO is an sp hybrid, different from their sp2hybrid arbitrarily chosen. We found that the lone-pair LMO at transition state for free acetylene can also be chosen as an sp hybrid, which is similar to that in the complexed form.
Table 2. Atomic Coordinates for 2f with Estimated Standard Deviations in Parentheses“ atom X Y 2 Ru -3579( 1) -1348(0) -6024(1) -3187(1) -7387(2) Br 1 -128(1) -3947(1) Br2 -2928( 1) -6035(2) -1 676(2) -1633(2) -4980(3) P1 -1074(2) P2 -7101(3) -5505(2) c1 -958( 14) -4361(15) -3582(11) c2 -3637( 17) -3024(22) -949( 13) c3 -1 75( 10) -3458( 12) -2093(13) c4 -139(11) -2328( 13) -936(15) -3400(25) 608(25) -3 140(28) c5 -4380(16) 87(21) -1627(20) C6 c11 -2658(8) -3958(13) ‘-1 243(9) c12 -1 708(10) -2856(10) -2903( 13) C13 -1348(13) -2022( 15) -3574(10) -2186( 16) -550(13) C14 -4097(10) -3917(10) -32 19(20) C15 -65(14) -3 189(9) -4117(16) C16 -445(12) -832( 10) -8 19(9) -3730( 13) c21 74( 10) -3987(15) -1064( 13) c22 682(10) -3039( 18) -395( 15) C23 C24 456( 17) 420(12) -1876(17) -1577(19) C25 672(14) -440( 12) 47(12) -25 11( 17) C26 -1062( 10) -1069(9) C31 -6362(12) -1721(8) -75( 10) -6198( 13) -1377(12) C32 c33 357(12) -7300( 17) -1446(12) c34 -8569(15) -1885(9) -257(13) c35 -1251( 12) -87 13(14) -2220( 12) -2152( 11) C36 -1663( 11) .-7613( 13) -6228(9) -1 154(9) C41 -5860( 13) 442(11) -5506(17) -6751( 13) C42 c43 -7295( 14) -546( 13) -4506( 18) c44 -3865( 16) -7281 (12) -1356( 16) c45 -6846(13) -4302(17) -2053( 14) -6275( 13) -5250( 15) C46 -1963(11) -6264(9) -8549( 12) C51 -1826(8) -5726(10) -2 155( 12) -9430( 13) C52 -6283( 12) -10568( 15) -2707( 11) c53 -7382( 12) -2925(10) -1 0837(15 ) c54 -9943(15) -7891(11) -2584(10) c55 -7338( 10) -2025(9) C56 -883 1(13) -5852(9) -798 1(12) C61 2(8) -6332(12) 86(10) -9439( 14) C62 -65 14(12) 927(10) -1OO68( 16) C63 -62 14(12) 1675(11) -9191(18) C64 -5747(12) 1592(9) -7814( 19) C65 -71 15(14) 746(8) -5529( 11) C66
lone-pair obital by accepting lone-pair electrons into a vacant metal d orbital. Finally, the presence of a six-coordiante alkynyl(hydrid0)ruthenium(1V) intermediate in the present tautomerization reaction has been examined. If the oxidative addition of the acetylenic C-H takes place readily in the present system, the possibility of Antonova’s 1,3-hydrogen shifts or Werner’s bimolecular hydrogen transferlh has to be considered. Since an octahedral d4 complex is expected to have two singly occupied d orbitals, the closed-shell MP2 calculations cannot be applied here. Assuming conservation of the spin state, we optimized the geometry of the oxidative addition product derived from C,i.e. RuCl,(PH3)2(H)((=-cH) (C)having a trans-dichloride and a cis-bisl(phosphine), with singlet UHF calculationsto a stationary
Behavior of Acetylene in a Coordination Sphere
J. Am. Chem. SOC.,Vol. 116, No. 18, 1994 8111
Table 3. Selected Interatomic Distances (A) and Angles (deg) for 2f Distances Ru-Br 1 Ru-P1 Ru-C 1 C2-C3 C3-a Br 1-Ru-Br2 Br 1-Ru-PI Br2-Ru-P 1 C1-Ru-Brl C 1-Ru-PI Ru-Cl-CZ
2.469(2) 2.384(3) 1.768( 17) 1.47(2) 1.61(4) Angles 144.8l(7) 89.3(1) 89.8(1) 109.6(7) 90.0(5) 161(2)
Ru-Br2 Ru-P~ C1-C2 c3-C4 C346 Pl-Ru-P2 Brl-Ru-P2 Br2-Ru-P2 C1-Ru-Br2 Cl-Ru-P2 Cl-C2-C3
2.460(2) 2.410(3) 1.36(3) 1.53(2) 1.50(3) 179.3( 1) 90.8( 1) 89.7(1) 105.6(7) 90.6(5) 125(2)
state shown in Figure 3. Due to the strong trans effect of the hydride ligand, the phoshpine trans to it has a long metal-P distance of 2.8 A. The energy was refined by theprojectedUMP2 method. To compare this energy level with the other complexes described above, we chose [TS2], the highest point so far, as a reference and reoptimized its structure with the RHF level and then refined the energy with the MP2 calculation on the RHF optimized structure. The result indicated that the oxidative addition product G is very unstable, its UMP2 energy being ca. 67.5 kcal/mol higher than the MP2 energy of [TSz]. Though the method of calculations used in the geometry optimization of C is different from others and hence exact comparison is not possible, it is qualitatively certain that the oxidative addition in the present system is thermodynamically very unfavorable, as indicated in Figure 4. The key complex C is the turning point of the reaction pathway, toward 1,Zhydrogen migration or to oxidative addition, the latter process being difficult if the metal changes from d6 to d4, as in the present case. For the d8 to d6 process, the oxidative addition should be just as easy as in the reaction of alkynes26 or alkane$' with Rh(1) Wilkinson-type complexes. (26) Wakatsuki, Y.;Koga, N.; Morokuma, K. Unpublished results.
Experimetnal Section IR spectra wereobtained with a Shimadzu IR-27G spectrometerusing the KBr pellet method. NMR spectra were recorded on a JEOL JNMGX-400 or GX-500 spectrometer using SiMe4 ('H) or H3P04 ("P) as internal standard. The vinylidene complexes, If and 2f,were prepared as described pre~iously.~ Single crystals of 2fsuitable for X-ray analysis were obtained by recrystallization from 1,2-dichloroethane/hexane. X-ray Crystal Structure Determination of 2f. Crystal data are summarized in Table 1. X-ray measurements were carried out with a Nonius CAD4 four-circle diffractometer equipped with a graphite monochromator using (u-28 scans. Absorption correction was not made because deviations of Fofor axial reflectionsat x = 90° were within 45%. A total of 4712 unique reflections in the range i h , ik, +I and 2O < 28 < 55' were measured, of which 3956 independent reflections having I > 3 4 ) were used in subsequent analysis. The structure was solved from direct and Fourier methods and refined by block-diagonal least squares with anisotropic thermal parameters in the last cycles for all non-hydrqgen atoms. Hydrogen atom8 for the six phenyl rings and the vinylic proton were placed in calculated positions, while those for the methyl groups were not located. In the refinements, unit weight was applied. The function minimized in the least-squares refinementwas Zw(!.Fd-IFd)Z. The handednessofthecrystalwascorrectly established, since R = 0.061 for t8e opposite handedness. The computational program package used in the analysis was a UNICS 3 program system.% Neutral atomic scattering factors were taken from the International Tables for X-ray crystallography.29 Final atomic parameters for the non-hydrogen atoms and important bond lengths and angles are given in Table 2 and 3, respectively.
Supplementary Material Available: Tables of positional and thermal parameters for 2f,a full tabulation of bond distances and angles, and Cartesian coordinates of atom positions calculated for A-F, [TSJ, and [TS2] (9 pages); a listing of observed and calculated structure factors (10 pages). This material is contained in many libraries on microfiche, immediately follows this article in the microfilm version of the journal, and can be ordered from the ACS; see any current masthead page for ordering information. (27) (a) Koga, N.; Morokuma, K. J . Phys. Chem. 1990.94,5454; (b) J. Am. Chem. Soc. 1993,115,6883. (28) Sakurai, T.; Kobayashi, K.Rikagaku Kenkyusho Hokoku 1979,55, 69. (29) International Tablesfor X-Ray Crystallography; Kynoch: Birmingham, England, 1976; Vol. 3, p 13.
